1 .. This work is licensed under a Creative Commons Attribution 4.0 International License.
2 .. http://creativecommons.org/licenses/by/4.0
3 .. (c) OPNFV, Intel Corporation, AT&T and others.
5 vSwitchPerf test suites userguide
6 ---------------------------------
11 VSPERF requires a traffic generators to run tests, automated traffic gen
12 support in VSPERF includes:
14 - IXIA traffic generator (IxNetwork hardware) and a machine that runs the IXIA
16 - Spirent traffic generator (TestCenter hardware chassis or TestCenter virtual
17 in a VM) and a VM to run the Spirent Virtual Deployment Service image,
18 formerly known as "Spirent LabServer".
19 - Xena Network traffic generator (Xena hardware chassis) that houses the Xena
20 Traffic generator modules.
21 - Moongen software traffic generator. Requires a separate machine running
22 moongen to execute packet generation.
24 If you want to use another traffic generator, please select the :ref:`trafficgen-dummy`
30 To see the supported Operating Systems, vSwitches and system requirements,
31 please follow the `installation instructions <vsperf-installation>`.
33 Traffic Generator Setup
34 ^^^^^^^^^^^^^^^^^^^^^^^
36 Follow the `Traffic generator instructions <trafficgen-installation>` to
37 install and configure a suitable traffic generator.
39 Cloning and building src dependencies
40 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
42 In order to run VSPERF, you will need to download DPDK and OVS. You can
43 do this manually and build them in a preferred location, OR you could
44 use vswitchperf/src. The vswitchperf/src directory contains makefiles
45 that will allow you to clone and build the libraries that VSPERF depends
46 on, such as DPDK and OVS. To clone and build simply:
48 .. code-block:: console
53 VSPERF can be used with stock OVS (without DPDK support). When build
54 is finished, the libraries are stored in src_vanilla directory.
56 The 'make' builds all options in src:
59 * OVS with vhost_user as the guest access method (with DPDK support)
61 The vhost_user build will reside in src/ovs/
62 The Vanilla OVS build will reside in vswitchperf/src_vanilla
64 To delete a src subdirectory and its contents to allow you to re-clone simply
67 .. code-block:: console
71 Configure the ``./conf/10_custom.conf`` file
72 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
74 The ``10_custom.conf`` file is the configuration file that overrides
75 default configurations in all the other configuration files in ``./conf``
76 The supplied ``10_custom.conf`` file **MUST** be modified, as it contains
77 configuration items for which there are no reasonable default values.
79 The configuration items that can be added is not limited to the initial
80 contents. Any configuration item mentioned in any .conf file in
81 ``./conf`` directory can be added and that item will be overridden by
82 the custom configuration value.
84 Further details about configuration files evaluation and special behaviour
85 of options with ``GUEST_`` prefix could be found at :ref:`design document
86 <design-configuration>`.
88 Using a custom settings file
89 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
91 If your ``10_custom.conf`` doesn't reside in the ``./conf`` directory
92 of if you want to use an alternative configuration file, the file can
93 be passed to ``vsperf`` via the ``--conf-file`` argument.
95 .. code-block:: console
97 $ ./vsperf --conf-file <path_to_custom_conf> ...
99 Note that configuration passed in via the environment (``--load-env``)
100 or via another command line argument will override both the default and
101 your custom configuration files. This "priority hierarchy" can be
102 described like so (1 = max priority):
104 1. Testcase definition section ``Parameters``
105 2. Command line arguments
106 3. Environment variables
107 4. Configuration file(s)
109 Further details about configuration files evaluation and special behaviour
110 of options with ``GUEST_`` prefix could be found at :ref:`design document
111 <design-configuration>`.
113 .. _overriding-parameters-documentation:
115 Referencing parameter values
116 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
118 It is possible to use a special macro ``#PARAM()`` to refer to the value of
119 another configuration parameter. This reference is evaluated during
120 access of the parameter value (by ``settings.getValue()`` call), so it
121 can refer to parameters created during VSPERF runtime, e.g. NICS dictionary.
122 It can be used to reflect DUT HW details in the testcase definition.
134 # set destination MAC to the MAC of the first
135 # interface from WHITELIST_NICS list
136 'dstmac' : '#PARAM(NICS[0]["mac"])',
141 Overriding values defined in configuration files
142 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
144 The configuration items can be overridden by command line argument
145 ``--test-params``. In this case, the configuration items and
146 their values should be passed in form of ``item=value`` and separated
153 $ ./vsperf --test-params "TRAFFICGEN_DURATION=10;TRAFFICGEN_PKT_SIZES=(128,);" \
154 "GUEST_LOOPBACK=['testpmd','l2fwd']" pvvp_tput
156 The second option is to override configuration items by ``Parameters`` section
157 of the test case definition. The configuration items can be added into ``Parameters``
158 dictionary with their new values. These values will override values defined in
159 configuration files or specified by ``--test-params`` command line argument.
165 "Parameters" : {'TRAFFICGEN_PKT_SIZES' : (128,),
166 'TRAFFICGEN_DURATION' : 10,
167 'GUEST_LOOPBACK' : ['testpmd','l2fwd'],
170 **NOTE:** In both cases, configuration item names and their values must be specified
171 in the same form as they are defined inside configuration files. Parameter names
172 must be specified in uppercase and data types of original and new value must match.
173 Python syntax rules related to data types and structures must be followed.
174 For example, parameter ``TRAFFICGEN_PKT_SIZES`` above is defined as a tuple
175 with a single value ``128``. In this case trailing comma is mandatory, otherwise
176 value can be wrongly interpreted as a number instead of a tuple and vsperf
177 execution would fail. Please check configuration files for default values and their
178 types and use them as a basis for any customized values. In case of any doubt, please
179 check official python documentation related to data structures like tuples, lists
182 **NOTE:** Vsperf execution will terminate with runtime error in case, that unknown
183 parameter name is passed via ``--test-params`` CLI argument or defined in ``Parameters``
184 section of test case definition. It is also forbidden to redefine a value of
185 ``TEST_PARAMS`` configuration item via CLI or ``Parameters`` section.
190 VSPERF uses a VM image called vloop_vnf for looping traffic in the deployment
191 scenarios involving VMs. The image can be downloaded from
192 `<http://artifacts.opnfv.org/>`__.
194 Please see the installation instructions for information on :ref:`vloop-vnf`
202 A Kernel Module that provides OSI Layer 2 Ipv4 termination or forwarding with
203 support for Destination Network Address Translation (DNAT) for both the MAC and
204 IP addresses. l2fwd can be found in <vswitchperf_dir>/src/l2fwd
209 All examples inside these docs assume, that user is inside the VSPERF
210 directory. VSPERF can be executed from any directory.
212 Before running any tests make sure you have root permissions by adding
213 the following line to /etc/sudoers:
215 .. code-block:: console
217 username ALL=(ALL) NOPASSWD: ALL
219 username in the example above should be replaced with a real username.
221 To list the available tests:
223 .. code-block:: console
227 To run a single test:
229 .. code-block:: console
233 Where $TESTNAME is the name of the vsperf test you would like to run.
235 To run a group of tests, for example all tests with a name containing
238 .. code-block:: console
240 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf --tests="RFC2544"
244 .. code-block:: console
246 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf
248 Some tests allow for configurable parameters, including test duration
249 (in seconds) as well as packet sizes (in bytes).
253 $ ./vsperf --conf-file user_settings.py \
254 --tests RFC2544Tput \
255 --test-params "TRAFFICGEN_DURATION=10;TRAFFICGEN_PKT_SIZES=(128,)"
257 For all available options, check out the help dialog:
259 .. code-block:: console
263 Executing Vanilla OVS tests
264 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
266 1. If needed, recompile src for all OVS variants
268 .. code-block:: console
274 2. Update your ``10_custom.conf`` file to use Vanilla OVS:
276 .. code-block:: python
278 VSWITCH = 'OvsVanilla'
282 .. code-block:: console
284 $ ./vsperf --conf-file=<path_to_custom_conf>
286 Please note if you don't want to configure Vanilla OVS through the
287 configuration file, you can pass it as a CLI argument.
289 .. code-block:: console
291 $ ./vsperf --vswitch OvsVanilla
294 Executing tests with VMs
295 ^^^^^^^^^^^^^^^^^^^^^^^^
297 To run tests using vhost-user as guest access method:
299 1. Set VSWITCH and VNF of your settings file to:
301 .. code-block:: python
303 VSWITCH = 'OvsDpdkVhost'
304 VNF = 'QemuDpdkVhost'
306 2. If needed, recompile src for all OVS variants
308 .. code-block:: console
316 .. code-block:: console
318 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf
320 **NOTE:** By default vSwitch is acting as a server for dpdk vhost-user sockets.
321 In case, that QEMU should be a server for vhost-user sockets, then parameter
322 ``VSWITCH_VHOSTUSER_SERVER_MODE`` should be set to ``False``.
324 Executing tests with VMs using Vanilla OVS
325 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
327 To run tests using Vanilla OVS:
329 1. Set the following variables:
331 .. code-block:: python
333 VSWITCH = 'OvsVanilla'
334 VNF = 'QemuVirtioNet'
336 VANILLA_TGEN_PORT1_IP = n.n.n.n
337 VANILLA_TGEN_PORT1_MAC = nn:nn:nn:nn:nn:nn
339 VANILLA_TGEN_PORT2_IP = n.n.n.n
340 VANILLA_TGEN_PORT2_MAC = nn:nn:nn:nn:nn:nn
342 VANILLA_BRIDGE_IP = n.n.n.n
344 or use ``--test-params`` option
346 .. code-block:: console
348 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf \
349 --test-params "VANILLA_TGEN_PORT1_IP=n.n.n.n;" \
350 "VANILLA_TGEN_PORT1_MAC=nn:nn:nn:nn:nn:nn;" \
351 "VANILLA_TGEN_PORT2_IP=n.n.n.n;" \
352 "VANILLA_TGEN_PORT2_MAC=nn:nn:nn:nn:nn:nn"
354 2. If needed, recompile src for all OVS variants
356 .. code-block:: console
364 .. code-block:: console
366 $ ./vsperf --conf-file<path_to_custom_conf>/10_custom.conf
373 Currently it is not possible to use standard scenario deployments for execution of
374 tests with VPP. It means, that deployments ``p2p``, ``pvp``, ``pvvp`` and in general any
375 :ref:`pxp-deployment` won't work with VPP. However it is possible to use VPP in
376 :ref:`step-driven-tests`. A basic set of VPP testcases covering ``phy2phy``, ``pvp``
377 and ``pvvp`` tests are already prepared.
379 List of performance tests with VPP support follows:
381 * phy2phy_tput_vpp: VPP: LTD.Throughput.RFC2544.PacketLossRatio
382 * phy2phy_cont_vpp: VPP: Phy2Phy Continuous Stream
383 * phy2phy_back2back_vpp: VPP: LTD.Throughput.RFC2544.BackToBackFrames
384 * pvp_tput_vpp: VPP: LTD.Throughput.RFC2544.PacketLossRatio
385 * pvp_cont_vpp: VPP: PVP Continuous Stream
386 * pvp_back2back_vpp: VPP: LTD.Throughput.RFC2544.BackToBackFrames
387 * pvvp_tput_vpp: VPP: LTD.Throughput.RFC2544.PacketLossRatio
388 * pvvp_cont_vpp: VPP: PVP Continuous Stream
389 * pvvp_back2back_vpp: VPP: LTD.Throughput.RFC2544.BackToBackFrames
391 In order to execute testcases with VPP it is required to:
393 * install VPP manually, see :ref:`vpp-installation`
394 * configure ``WHITELIST_NICS``, with two physical NICs connected to the traffic generator
395 * configure traffic generator, see :ref:`trafficgen-installation`
397 After that it is possible to execute VPP testcases listed above.
401 .. code-block:: console
403 $ ./vsperf --conf-file=<path_to_custom_conf> phy2phy_tput_vpp
407 Using vfio_pci with DPDK
408 ^^^^^^^^^^^^^^^^^^^^^^^^^
410 To use vfio with DPDK instead of igb_uio add into your custom configuration
411 file the following parameter:
413 .. code-block:: python
415 PATHS['dpdk']['src']['modules'] = ['uio', 'vfio-pci']
418 **NOTE:** In case, that DPDK is installed from binary package, then please
419 set ``PATHS['dpdk']['bin']['modules']`` instead.
421 **NOTE:** Please ensure that Intel VT-d is enabled in BIOS.
423 **NOTE:** Please ensure your boot/grub parameters include
426 **NOTE:** In case of VPP, it is required to explicitly define, that vfio-pci
427 DPDK driver should be used. It means to update dpdk part of VSWITCH_VPP_ARGS
428 dictionary with uio-driver section, e.g. VSWITCH_VPP_ARGS['dpdk'] = 'uio-driver vfio-pci'
430 .. code-block:: console
432 iommu=pt intel_iommu=on
434 To check that IOMMU is enabled on your platform:
436 .. code-block:: console
439 [ 0.000000] Intel-IOMMU: enabled
440 [ 0.139882] dmar: IOMMU 0: reg_base_addr fbffe000 ver 1:0 cap d2078c106f0466 ecap f020de
441 [ 0.139888] dmar: IOMMU 1: reg_base_addr ebffc000 ver 1:0 cap d2078c106f0466 ecap f020de
442 [ 0.139893] IOAPIC id 2 under DRHD base 0xfbffe000 IOMMU 0
443 [ 0.139894] IOAPIC id 0 under DRHD base 0xebffc000 IOMMU 1
444 [ 0.139895] IOAPIC id 1 under DRHD base 0xebffc000 IOMMU 1
445 [ 3.335744] IOMMU: dmar0 using Queued invalidation
446 [ 3.335746] IOMMU: dmar1 using Queued invalidation
454 To use virtual functions of NIC with SRIOV support, use extended form
455 of NIC PCI slot definition:
457 .. code-block:: python
459 WHITELIST_NICS = ['0000:05:00.0|vf0', '0000:05:00.1|vf3']
461 Where 'vf' is an indication of virtual function usage and following
462 number defines a VF to be used. In case that VF usage is detected,
463 then vswitchperf will enable SRIOV support for given card and it will
464 detect PCI slot numbers of selected VFs.
466 So in example above, one VF will be configured for NIC '0000:05:00.0'
467 and four VFs will be configured for NIC '0000:05:00.1'. Vswitchperf
468 will detect PCI addresses of selected VFs and it will use them during
471 At the end of vswitchperf execution, SRIOV support will be disabled.
473 SRIOV support is generic and it can be used in different testing scenarios.
476 * vSwitch tests with DPDK or without DPDK support to verify impact
477 of VF usage on vSwitch performance
478 * tests without vSwitch, where traffic is forwared directly
479 between VF interfaces by packet forwarder (e.g. testpmd application)
480 * tests without vSwitch, where VM accesses VF interfaces directly
481 by PCI-passthrough_ to measure raw VM throughput performance.
485 Using QEMU with PCI passthrough support
486 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
488 Raw virtual machine throughput performance can be measured by execution of PVP
489 test with direct access to NICs by PCI passthrough. To execute VM with direct
490 access to PCI devices, enable vfio-pci_. In order to use virtual functions,
491 SRIOV-support_ must be enabled.
493 Execution of test with PCI passthrough with vswitch disabled:
495 .. code-block:: console
497 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf \
498 --vswitch none --vnf QemuPciPassthrough pvp_tput
500 Any of supported guest-loopback-application_ can be used inside VM with
501 PCI passthrough support.
503 Note: Qemu with PCI passthrough support can be used only with PVP test
506 .. _guest-loopback-application:
508 Selection of loopback application for tests with VMs
509 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
511 To select the loopback applications which will forward packets inside VMs,
512 the following parameter should be configured:
514 .. code-block:: python
516 GUEST_LOOPBACK = ['testpmd']
518 or use ``--test-params`` CLI argument:
520 .. code-block:: console
522 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf \
523 --test-params "GUEST_LOOPBACK=['testpmd']"
525 Supported loopback applications are:
527 .. code-block:: console
529 'testpmd' - testpmd from dpdk will be built and used
530 'l2fwd' - l2fwd module provided by Huawei will be built and used
531 'linux_bridge' - linux bridge will be configured
532 'buildin' - nothing will be configured by vsperf; VM image must
533 ensure traffic forwarding between its interfaces
535 Guest loopback application must be configured, otherwise traffic
536 will not be forwarded by VM and testcases with VM related deployments
537 will fail. Guest loopback application is set to 'testpmd' by default.
539 **NOTE:** In case that only 1 or more than 2 NICs are configured for VM,
540 then 'testpmd' should be used. As it is able to forward traffic between
541 multiple VM NIC pairs.
543 **NOTE:** In case of linux_bridge, all guest NICs are connected to the same
544 bridge inside the guest.
546 Mergable Buffers Options with QEMU
547 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
549 Mergable buffers can be disabled with VSPerf within QEMU. This option can
550 increase performance significantly when not using jumbo frame sized packets.
551 By default VSPerf disables mergable buffers. If you wish to enable it you
552 can modify the setting in the a custom conf file.
554 .. code-block:: python
556 GUEST_NIC_MERGE_BUFFERS_DISABLE = [False]
558 Then execute using the custom conf file.
560 .. code-block:: console
562 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf
564 Alternatively you can just pass the param during execution.
566 .. code-block:: console
568 $ ./vsperf --test-params "GUEST_NIC_MERGE_BUFFERS_DISABLE=[False]"
571 Selection of dpdk binding driver for tests with VMs
572 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
574 To select dpdk binding driver, which will specify which driver the vm NICs will
575 use for dpdk bind, the following configuration parameter should be configured:
577 .. code-block:: console
579 GUEST_DPDK_BIND_DRIVER = ['igb_uio_from_src']
581 The supported dpdk guest bind drivers are:
583 .. code-block:: console
585 'uio_pci_generic' - Use uio_pci_generic driver
586 'igb_uio_from_src' - Build and use the igb_uio driver from the dpdk src
588 'vfio_no_iommu' - Use vfio with no iommu option. This requires custom
589 guest images that support this option. The default
590 vloop image does not support this driver.
592 Note: uio_pci_generic does not support sr-iov testcases with guests attached.
593 This is because uio_pci_generic only supports legacy interrupts. In case
594 uio_pci_generic is selected with the vnf as QemuPciPassthrough it will be
595 modified to use igb_uio_from_src instead.
597 Note: vfio_no_iommu requires kernels equal to or greater than 4.5 and dpdk
598 16.04 or greater. Using this option will also taint the kernel.
600 Please refer to the dpdk documents at http://dpdk.org/doc/guides for more
601 information on these drivers.
603 Guest Core and Thread Binding
604 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
606 VSPERF provides options to achieve better performance by guest core binding and
607 guest vCPU thread binding as well. Core binding is to bind all the qemu threads.
608 Thread binding is to bind the house keeping threads to some CPU and vCPU thread to
609 some other CPU, this helps to reduce the noise from qemu house keeping threads.
612 .. code-block:: python
614 GUEST_CORE_BINDING = [('#EVAL(6+2*#VMINDEX)', '#EVAL(7+2*#VMINDEX)')]
616 **NOTE** By default the GUEST_THREAD_BINDING will be none, which means same as
617 the GUEST_CORE_BINDING, i.e. the vcpu threads are sharing the physical CPUs with
618 the house keeping threads. Better performance using vCPU thread binding can be
619 achieved by enabling affinity in the custom configuration file.
621 For example, if an environment requires 32,33 to be core binded and 29,30&31 for
622 guest thread binding to achieve better performance.
624 .. code-block:: python
626 VNF_AFFINITIZATION_ON = True
627 GUEST_CORE_BINDING = [('32','33')]
628 GUEST_THREAD_BINDING = [('29', '30', '31')]
633 QEMU default to a compatible subset of performance enhancing cpu features.
634 To pass all available host processor features to the guest.
636 .. code-block:: python
638 GUEST_CPU_OPTIONS = ['host,migratable=off']
640 **NOTE** To enhance the performance, cpu features tsc deadline timer for guest,
641 the guest PMU, the invariant TSC can be provided in the custom configuration file.
643 Multi-Queue Configuration
644 ^^^^^^^^^^^^^^^^^^^^^^^^^
646 VSPerf currently supports multi-queue with the following limitations:
648 1. Requires QEMU 2.5 or greater and any OVS version higher than 2.5. The
649 default upstream package versions installed by VSPerf satisfies this
652 2. Guest image must have ethtool utility installed if using l2fwd or linux
653 bridge inside guest for loopback.
655 3. If using OVS versions 2.5.0 or less enable old style multi-queue as shown
656 in the ''02_vswitch.conf'' file.
658 .. code-block:: python
660 OVS_OLD_STYLE_MQ = True
662 To enable multi-queue for dpdk modify the ''02_vswitch.conf'' file.
664 .. code-block:: python
666 VSWITCH_DPDK_MULTI_QUEUES = 2
668 **NOTE:** you should consider using the switch affinity to set a pmd cpu mask
669 that can optimize your performance. Consider the numa of the NIC in use if this
670 applies by checking /sys/class/net/<eth_name>/device/numa_node and setting an
671 appropriate mask to create PMD threads on the same numa node.
673 When multi-queue is enabled, each dpdk or dpdkvhostuser port that is created
674 on the switch will set the option for multiple queues. If old style multi queue
675 has been enabled a global option for multi queue will be used instead of the
678 To enable multi-queue on the guest modify the ''04_vnf.conf'' file.
680 .. code-block:: python
682 GUEST_NIC_QUEUES = [2]
684 Enabling multi-queue at the guest will add multiple queues to each NIC port when
685 qemu launches the guest.
687 In case of Vanilla OVS, multi-queue is enabled on the tuntap ports and nic
688 queues will be enabled inside the guest with ethtool. Simply enabling the
689 multi-queue on the guest is sufficient for Vanilla OVS multi-queue.
691 Testpmd should be configured to take advantage of multi-queue on the guest if
692 using DPDKVhostUser. This can be done by modifying the ''04_vnf.conf'' file.
694 .. code-block:: python
696 GUEST_TESTPMD_PARAMS = ['-l 0,1,2,3,4 -n 4 --socket-mem 512 -- '
697 '--burst=64 -i --txqflags=0xf00 '
698 '--nb-cores=4 --rxq=2 --txq=2 '
701 **NOTE:** The guest SMP cores must be configured to allow for testpmd to use the
702 optimal number of cores to take advantage of the multiple guest queues.
704 In case of using Vanilla OVS and qemu virtio-net you can increase performance
705 by binding vhost-net threads to cpus. This can be done by enabling the affinity
706 in the ''04_vnf.conf'' file. This can be done to non multi-queue enabled
707 configurations as well as there will be 2 vhost-net threads.
709 .. code-block:: python
711 VSWITCH_VHOST_NET_AFFINITIZATION = True
713 VSWITCH_VHOST_CPU_MAP = [4,5,8,11]
715 **NOTE:** This method of binding would require a custom script in a real
718 **NOTE:** For optimal performance guest SMPs and/or vhost-net threads should be
719 on the same numa as the NIC in use if possible/applicable. Testpmd should be
720 assigned at least (nb_cores +1) total cores with the cpu mask.
725 VSPERF provides options to support jumbo frame testing with a jumbo frame supported
726 NIC and traffic generator for the following vswitches:
732 3. TestPMD loopback with or without a guest
734 **NOTE:** There is currently no support for SR-IOV or VPP at this time with jumbo
737 All packet forwarding applications for pxp testing is supported.
739 To enable jumbo frame testing simply enable the option in the conf files and set the
740 maximum size that will be used.
742 .. code-block:: python
744 VSWITCH_JUMBO_FRAMES_ENABLED = True
745 VSWITCH_JUMBO_FRAMES_SIZE = 9000
747 To enable jumbo frame testing with OVSVanilla the NIC in test on the host must have
748 its mtu size changed manually using ifconfig or applicable tools:
750 .. code-block:: console
752 ifconfig eth1 mtu 9000 up
754 **NOTE:** To make the setting consistent across reboots you should reference the OS
755 documents as it differs from distribution to distribution.
757 To start a test for jumbo frames modify the conf file packet sizes or pass the option
758 through the VSPERF command line.
760 .. code-block:: python
762 TEST_PARAMS = {'TRAFFICGEN_PKT_SIZES':(2000,9000)}
764 .. code-block:: python
766 ./vsperf --test-params "TRAFFICGEN_PKT_SIZES=2000,9000"
768 It is recommended to increase the memory size for OvsDpdkVhostUser testing from the default
769 1024. Your size required may vary depending on the number of guests in your testing. 4096
770 appears to work well for most typical testing scenarios.
772 .. code-block:: python
774 DPDK_SOCKET_MEM = ['4096', '0']
776 **NOTE:** For Jumbo frames to work with DpdkVhostUser, mergable buffers will be enabled by
777 default. If testing with mergable buffers in QEMU is desired, disable Jumbo Frames and only
778 test non jumbo frame sizes. Test Jumbo Frames sizes separately to avoid this collision.
781 Executing Packet Forwarding tests
782 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
784 To select the applications which will forward packets,
785 the following parameters should be configured:
787 .. code-block:: python
792 or use ``--vswitch`` and ``--fwdapp`` CLI arguments:
794 .. code-block:: console
796 $ ./vsperf phy2phy_cont --conf-file user_settings.py \
800 Supported Packet Forwarding applications are:
802 .. code-block:: console
804 'testpmd' - testpmd from dpdk
807 1. Update your ''10_custom.conf'' file to use the appropriate variables
808 for selected Packet Forwarder:
810 .. code-block:: python
812 # testpmd configuration
814 # packet forwarding mode supported by testpmd; Please see DPDK documentation
815 # for comprehensive list of modes supported by your version.
816 # e.g. io|mac|mac_retry|macswap|flowgen|rxonly|txonly|csum|icmpecho|...
817 # Note: Option "mac_retry" has been changed to "mac retry" since DPDK v16.07
818 TESTPMD_FWD_MODE = 'csum'
819 # checksum calculation layer: ip|udp|tcp|sctp|outer-ip
820 TESTPMD_CSUM_LAYER = 'ip'
821 # checksum calculation place: hw (hardware) | sw (software)
822 TESTPMD_CSUM_CALC = 'sw'
823 # recognize tunnel headers: on|off
824 TESTPMD_CSUM_PARSE_TUNNEL = 'off'
828 .. code-block:: console
830 $ ./vsperf phy2phy_tput --conf-file <path_to_settings_py>
832 Executing Packet Forwarding tests with one guest
833 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
835 TestPMD with DPDK 16.11 or greater can be used to forward packets as a switch to a single guest using TestPMD vdev
836 option. To set this configuration the following parameters should be used.
838 .. code-block:: python
843 or use ``--vswitch`` and ``--fwdapp`` CLI arguments:
845 .. code-block:: console
847 $ ./vsperf pvp_tput --conf-file user_settings.py \
851 Guest forwarding application only supports TestPMD in this configuration.
853 .. code-block:: python
855 GUEST_LOOPBACK = ['testpmd']
857 For optimal performance one cpu per port +1 should be used for TestPMD. Also set additional params for packet forwarding
858 application to use the correct number of nb-cores.
860 .. code-block:: python
862 DPDK_SOCKET_MEM = ['1024', '0']
863 VSWITCHD_DPDK_ARGS = ['-l', '46,44,42,40,38', '-n', '4']
864 TESTPMD_ARGS = ['--nb-cores=4', '--txq=1', '--rxq=1']
866 For guest TestPMD 3 VCpus should be assigned with the following TestPMD params.
868 .. code-block:: python
870 GUEST_TESTPMD_PARAMS = ['-l 0,1,2 -n 4 --socket-mem 1024 -- '
871 '--burst=64 -i --txqflags=0xf00 '
872 '--disable-hw-vlan --nb-cores=2 --txq=1 --rxq=1']
874 Execution of TestPMD can be run with the following command line
876 .. code-block:: console
878 ./vsperf pvp_tput --vswitch=none --fwdapp=TestPMD --conf-file <path_to_settings_py>
880 **NOTE:** To achieve the best 0% loss numbers with rfc2544 throughput testing, other tunings should be applied to host
881 and guest such as tuned profiles and CPU tunings to prevent possible interrupts to worker threads.
883 VSPERF modes of operation
884 ^^^^^^^^^^^^^^^^^^^^^^^^^
886 VSPERF can be run in different modes. By default it will configure vSwitch,
887 traffic generator and VNF. However it can be used just for configuration
888 and execution of traffic generator. Another option is execution of all
889 components except traffic generator itself.
891 Mode of operation is driven by configuration parameter -m or --mode
893 .. code-block:: console
895 -m MODE, --mode MODE vsperf mode of operation;
897 "normal" - execute vSwitch, VNF and traffic generator
898 "trafficgen" - execute only traffic generator
899 "trafficgen-off" - execute vSwitch and VNF
900 "trafficgen-pause" - execute vSwitch and VNF but wait before traffic transmission
902 In case, that VSPERF is executed in "trafficgen" mode, then configuration
903 of traffic generator can be modified through ``TRAFFIC`` dictionary passed to the
904 ``--test-params`` option. It is not needed to specify all values of ``TRAFFIC``
905 dictionary. It is sufficient to specify only values, which should be changed.
906 Detailed description of ``TRAFFIC`` dictionary can be found at
907 :ref:`configuration-of-traffic-dictionary`.
909 Example of execution of VSPERF in "trafficgen" mode:
911 .. code-block:: console
913 $ ./vsperf -m trafficgen --trafficgen IxNet --conf-file vsperf.conf \
914 --test-params "TRAFFIC={'traffic_type':'rfc2544_continuous','bidir':'False','framerate':60}"
916 Code change verification by pylint
917 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
919 Every developer participating in VSPERF project should run
920 pylint before his python code is submitted for review. Project
921 specific configuration for pylint is available at 'pylint.rc'.
923 Example of manual pylint invocation:
925 .. code-block:: console
927 $ pylint --rcfile ./pylintrc ./vsperf
932 Custom image fails to boot
933 ~~~~~~~~~~~~~~~~~~~~~~~~~~
935 Using custom VM images may not boot within VSPerf pxp testing because of
936 the drive boot and shared type which could be caused by a missing scsi
937 driver inside the image. In case of issues you can try changing the drive
940 .. code-block:: python
942 GUEST_BOOT_DRIVE_TYPE = ['ide']
943 GUEST_SHARED_DRIVE_TYPE = ['ide']
945 OVS with DPDK and QEMU
946 ~~~~~~~~~~~~~~~~~~~~~~~
948 If you encounter the following error: "before (last 100 chars):
949 '-path=/dev/hugepages,share=on: unable to map backing store for
950 hugepages: Cannot allocate memory\r\n\r\n" during qemu initialization,
951 check the amount of hugepages on your system:
953 .. code-block:: console
955 $ cat /proc/meminfo | grep HugePages
958 By default the vswitchd is launched with 1Gb of memory, to change
959 this, modify --socket-mem parameter in conf/02_vswitch.conf to allocate
960 an appropriate amount of memory:
962 .. code-block:: python
964 DPDK_SOCKET_MEM = ['1024', '0']
965 VSWITCHD_DPDK_ARGS = ['-c', '0x4', '-n', '4']
966 VSWITCHD_DPDK_CONFIG = {
967 'dpdk-init' : 'true',
968 'dpdk-lcore-mask' : '0x4',
969 'dpdk-socket-mem' : '1024,0',
972 Note: Option ``VSWITCHD_DPDK_ARGS`` is used for vswitchd, which supports ``--dpdk``
973 parameter. In recent vswitchd versions, option ``VSWITCHD_DPDK_CONFIG`` will be
974 used to configure vswitchd via ``ovs-vsctl`` calls.
980 For more information and details refer to the rest of vSwitchPerfuser documentation.